Specialize / Private Fine-Tuning
Hebbian specializes models only when the evidence says it is worth it.
We measure first. If a general model cannot meet the required judgment, we design the corpus, benchmark, and training with explicit limits.
Evaluation protocol
Train with evidence.
Hebbian compares models, reviews answers, and decides whether private training makes operating sense.
What the general model does today.
What quality, latency, or consistency is missing.
What difference would justify training.
Evidence before training
Benchmark before corpus.
That keeps training from becoming enthusiasm and focuses the work on visible improvement.
- Documented comparative benchmark.
- Private corpus with usage criteria.
- Honest reading of limits and results.
Lab method
Lab method.
Open a baseline with clear prompts and categories.
Prepare examples with intent and quality control.
Run training only when the gap justifies it.
Document findings, limits, and the next decision.
Success criteria
Proven improvement.
Answers improve in relevant categories.
The system responds within the useful margin.
Limits remain visible for the next decision.
What remains
Laboratory judgment.
Fine-tuning is treated as a business decision: if the gap does not show up in evidence, training is not forced.
The current model is measured before touching corpus or infrastructure.
Score, latency, loops, and limits are read together to avoid cosmetic conclusions.
The result clarifies whether to train, change approach, or strengthen the benchmark.
Hebbian Research Notes
Paper #001.
The Specialize page keeps the lab visible: benchmark, findings, limits, and a documented run.
Real artifact
The first gallery documents the LoRA run from April 24, 2026.
Next move
Train with evidence.
Hebbian specializes models when the benchmark proves there is a real improvement to capture.